SUMMARY We want to elucidate modes of polypeptide assembly that are important for biological function and associated with human disease but are difficult to characterize via standard experimental approaches. In each case, we wish to understand the non-covalent forces that underlie the assembly mode. Because of differences among the types of assembly we are studying, the experimental approaches we adopt are variable. One goal is to characterize quaternary structures formed by single-pass transmembrane (SPTM) ?-helices that are constituents of cell-surface receptors, such as receptor tyrosine kinases (RTKs) and activating immune receptors. These receptors form oligomers, and rearrangement of oligomer geometry transmits the signal. Crystallographic data are available in only for the SPTM ?-helix of immune receptor component DAP12. There is no high-resolution structural information for alternative geometries of SPTM ?-helix assemblies that are thought to be associated with different receptor activation states. We are applying racemic crystallization and micro-electron diffraction (via collaboration) to this structural challenge, and we are exploring protein-based ?picodiscs? as hosts for SPTM ?-helix assemblies. A second goal is to understand how sequence, composition and dimensions influence stability of polypeptides in the amyloid state. The ?-sheet-rich structures that are common to disease-associated amyloid fibrils are distinct from tertiary and quaternary structures commonly found among soluble proteins. The techniques commonly used to elucidate sequence-stability relationships among soluble proteins are not readily applied to amyloid fibrils; therefore, we are developing a soluble model system for the amyloid state, which we will exploit to ask fundamental questions about amyloid state stability. The third goal is to understand the forces that lead to liquid-liquid phase separation (LLPS) mediated by proteins in the FUS (?fused in sarcoma?) family. The loose associations between polypeptide chains in the protein-rich liquid phase are not well understood. Such phases can transition to amyloid-like assemblies, which forms a strong connection between our second and third goals. These more ordered assemblies are associated with illnesses such as ALS. We seek to model LLPS of FUS family proteins with synthetic peptides in order to conduct incisive tests of recent mechanistic proposals and to evaluate the role of amino acid sequence and stereochemistry in LLPS and the transition to more ordered and pathogenic assemblies.